Calculating a Dispatcher's Relative Share Based on Relative Queue Length and Capacity Value of a Plurality of Workload Types and Computing Systems Combinations

1. A method, in a workload manager, for distributing workload items to a plurality of dispatchers, each dispatcher being associated with a different computing system of a plurality of computing systems, the workload items comprising workload items of a plurality of different workload types, the method performed by the workload manager comprising: obtaining a capacity value for a combination of a given workload type and a given computing system in the plurality of computing systems, the capacity value representing a total capacity of workload items of the given workload type that the given computing system may process in a given time period; obtaining for the combination of the given workload type and the given computing system, a queue length of a given dispatcher associated with the given computing system; calculating, for the combination of the given workload type and the given computing system, a relative share of incoming workloads based on the queue length for the given dispatcher and the capacity value associated with the given workload type and the given computing system; and routing incoming workload items to the given dispatcher based on the calculated relative share for the given dispatcher.

2. The method of claim 1 , further comprising: obtaining, in the workload manager, for the combination of the given workload type and the given computing system, a total value of service units for the given workload type on the given computing system, wherein, for the combination of the given workload type and the given computing system, the relative share of incoming workloads is calculated based on the total value of service units for the given workload type on the given computing system.

3. The method of claim 2 , wherein the total value of service units for the combination of the given workload type and the given computing system, and the capacity value for the combination of the given workload type and the given computing system, are obtained from a memory associated with the workload manager.

4. The method of claim 1 , wherein calculating the relative share comprises: dividing the queue length of the given dispatcher by the capacity value for the combination of the given workload type and the given computing system.

5. The method of claim 1 , wherein the plurality of workload types comprises a general processor workload type (CP), an application assist processor workload type (zAAP), and an integrated information processor workload type (zIIP).

6. The method of claim 1 , wherein the plurality of computing systems are part of a sysplex.

7. The method of claim 1 , wherein routing the incoming workload items to the given dispatcher comprises sending a routing recommendation to an arbitrator.

8. The method of claim 7 , wherein the arbitrator distributes the incoming workload items to the given dispatcher according to the routing recommendation.

9. The method of claim 8 , wherein the given dispatcher receives the incoming workload items and queues the incoming workload items to be fetched by one or more execution units.

10. The method of claim 1 , wherein routing the incoming workload items to the given dispatcher comprises sending routing recommendations to an arbitrator, wherein the arbitrator distributes the incoming workload items to the plurality of dispatchers according to the routing recommendations.

11. A method, in a workload manager, for distributing workload items to a plurality of dispatchers, each dispatcher being associated with a different computing system of a plurality of computing systems, the workload items comprising workload items of a plurality of different workload types, the method performed by the workload manager comprising: calculating a capacity percentage for each combination of a given workload type and a given computing system in the plurality of computing systems, the capacity percentage representing a percentage of workload items of the given workload type that the given computing system can process in a given time period; determining a minimum capacity percentage from the calculated capacity percentage from each combination of workload types and computing systems; calculating, for each combination of a given workload type and a given computing system, a relative queue length of a dispatcher associated with the given computing system; determining a maximum relative queue length of a dispatcher from the calculated relative queue length from each combination of workload types and computing systems; determining, for each given computing system, a dispatcher's relative share of incoming workloads based on the minimum capacity percentage and the maximum relative queue length of the given computing system; and routing incoming workload items to the plurality of dispatchers based on the calculated relative share for each dispatcher.

12. The method of claim 11 , wherein calculating the capacity percentage for each combination of a given workload type and a given computing system comprises: obtaining, in the workload manager, a total value of service units of the given workload type across the plurality of computing systems; obtaining, in the workload manager, a capacity value for the given workload type and the given computing system, the capacity value representing a total capacity of workload items of the given workload type that the given computing system can process in a given time period; and dividing the capacity value of the given computing system for the given workload type by the total value of service units of the given workload type to form the capacity percentage.

13. The method of claim 11 , wherein calculating the relative queue length for each combination of a given workload type and a given computing system comprises: obtaining, in the workload manager, a total value of service units of the given workload type across the plurality of computing systems; obtaining, in the workload manager, a queue length for the given workload type and the given computing system; and dividing the queue length for the given workload type and the given computing system by the total value of service units of the given workload type across the plurality of computing systems.

14. The method of claim 11 , wherein determining a dispatcher's relative share for each given computing system comprises: determining an inverse of one minus the maximum relative queue length of the given computing system to form an intermediate result; and multiplying the minimum capacity percentage of the given computing system by the intermediate result to form the relative share.

15. The method of claim 11 , wherein the plurality of workload types comprises a general processor workload type (CP), an application assist processor workload type (zAAP), and an integrated information processor workload type (zIIP).

16. The method of claim 11 , wherein the plurality of computing systems are part of a sysplex.

17. The method of claim 11 , wherein routing the incoming workload items to the plurality of dispatchers comprises sending routing recommendations to one or more arbitrators.

18. The method of claim 17 , wherein the one or more arbitrators distribute the incoming workload items to the plurality of dispatchers according to the routing recommendations.

19. The method of claim 18 , wherein the plurality of dispatchers receive the incoming workload items and queue the incoming workload items to be fetched by a plurality of execution units.

20. The method of claim 11 , wherein routing the incoming workload items to the plurality of dispatchers comprises sending routing recommendations to an arbitrator, wherein the arbitrator distributes the incoming workload items to the plurality of dispatchers according to the routing recommendations.